The Inefficient Market Hypothesis

The effi­cient-mar­ket hy­poth­e­sis (EMH) is the idea that there are no hun­dred-dol­lar bills ly­ing on the side­walk be­cause some­one smarter than you would have picked them up by now. The EMH is a good tool for most peo­ple.

If you’re smart enough then you should re­verse this ad­vice into “There are hun­dred-dol­lar bills ly­ing on the side­walk”. If you are a ge­nius then you should re­verse it to the ex­treme. “There are hun­dred-dol­lar bills[1] ly­ing around all over the place.”

Hun­dred-dol­lar bills ly­ing on the side­walk are called “alpha”. Alpha is a tau­tolog­i­cally self-keep­ing se­cret. You can’t read about it in books. You can’t find it on blogs[2]. You will never be taught about it in school. “You can only find out about [alpha] if you go look­ing for it and you’ll only go look­ing for it if you already know it ex­ists.”

Where should you look?

Abstractions

A sys­tem is only as se­cure as its weak­est link. Crack­ing a sys­tem tends to hap­pen on an over­looked layer of ab­strac­tion[3].

  • It’s eas­ier to in­stall a key­log­ger than to break a good cryp­to­graphic pro­to­col.

  • It’s eas­ier to dis­assem­ble a com­puter and read the hard drive di­rectly[4] than to crack some­one’s pass­word.

The best at­tacks (those re­quiring the least work) hap­pen on an sep­a­rate di­men­sion of or­thog­o­nal­ity en­tirely.

  • The eas­iest way to talk to some­one pow­er­ful is just to call zir com­pany and ask by first name[5].

Won’t this tech­nique stop work­ing now that Tim Fer­ris has pub­lished it in a best­sel­ling book? Not nec­es­sar­ily. Quan­tum me­chan­ics has been pub­lic knowl­edge for decades yet most peo­ple can’t do it. The hard part of pick­pock­et­ing isn’t find­ing pock­ets to pick.

Per­haps you don’t need to talk to any­one rich and pow­er­ful. That is a good prob­lem to have.

I think you should find a prob­lem that’s easy for you to solve. Op­ti­miz­ing in solu­tion-space is fa­mil­iar and straight­for­ward, but you can make enor­mous gains play­ing around in prob­lem-space.

What Star­tups Are Really Like by Paul Graham

Prob­lem-space tends to have higher di­men­sion­al­ity than solu­tion space.

Case study

Ac­cord­ing to Joel Spolsky, the best pro­gram­mers have the abil­ity “to think in ab­strac­tions, and, most im­por­tantly, to view a prob­lem at sev­eral lev­els of ab­strac­tion si­mul­ta­neously.” Also ac­cord­ing to Joel Spolsky, a busi­ness is an “ab­strac­tion [that] ex­ists solely to cre­ate the illu­sion that the daily ac­tivi­ties of a pro­gram­mer (de­sign and writ­ing code, check­ing in code, de­bug­ging, etc.) are all that it takes to cre­ate soft­ware prod­ucts and bring them to mar­ket.”

The ideal pro­gram­mer em­ployee is some­one who can see all the way down to the level of bits, yet can’t raise zir head high enough to ma­nipu­late the fi­nan­cial ma­chin­ery of ven­ture cap­i­tal.

Home­work as­sign­ment: How can you har­vest alpha from this lo­cal equil­ibrium?

How to tell when you get it right

Alpha of­ten feels like a magic trick. You know the phrase “A ma­gi­cian never re­veals his se­crets”? Magic se­crets are not se­cret. The Ma­gi­cian’s Oath is not real. David Cop­perfield patents his in­ven­tions[6]. You can look them up in a pub­lic gov­ern­ment reg­istry. You don’t be­cause mag­i­cal se­crets are bor­ing. Dis­ap­point­ingly so.

Ma­gi­ci­ans cheat. The purest alpha should feel like cheat­ing too. The great­est com­pli­ment you can re­ceive about your alpha source isn’t “You’re a ge­nius.” It’s “That shouldn’t be pos­si­ble. I’m dis­illu­sioned to live in a world is so in­effi­cient.”

Of course, you should never hear ei­ther re­sponse be­cause you should never flaunt these dis­cov­er­ies in the first place.


  1. A month ago I offered to put Western­ers in touch with an N95 mask ex­porter in China. Only two read­ers took the effort to mes­sage me about it. One of them couldn’t be both­ered to use WeChat. ↩︎

  2. Ac­tu­ally, I did find alpha on a blog post once. The tu­to­rial has since been taken down. ↩︎

  3. For prac­tice, check out What is the fastest you can sort a list of ints in Java? ↩︎

  4. Most com­put­ers are not en­crypted. Pro­fes­sional soft­ware en­g­ineers are con­sistenly sur­prised by my abil­ity to re­cover files from a bro­ken lap­top with­out their lo­gin in­for­ma­tion. ↩︎

  5. Tim Fer­ris claims this works. I am in­clined to be­lieve him based on the guests who have at­tended his pod­cast. ↩︎

  6. Edit: I can only find one patent in­vented by David Cop­perfield, patent num­ber 90171779358477. Most of his patentable illu­sions seem to be in­vented by other peo­ple. ↩︎